ηk-nearest neighbor algorithm for estimation of symbolic user location in pervasive computing environments

The paper introduces a novel algorithm for the location-awareness problem of estimating symbolic user location for indoor spaces using IEEE 802.11 (Wi-Fi) wireless signals. The characteristic of the problem is that the signals fluctuate greatly, not only across perturbations in space, but also in time (diurnally), which leads to poor location estimation. The ηk-nearest neighbour (ηk-NN) algorithm is an instance-based learning algorithm which normalizes the sample data set of the Wi-Fi signal strength and signal quality to achieve the maximum correct result of symbolic user location at a room scale. Data normalization is found to play an important role in determining the quality of the training data-set which has direct impact on the estimation result. The algorithm has been compared to other k-nearest neighbour (k-NN) and shows promising results.